InclusionAILanguage modelOpen source

Ming-flash-omni-2.0

Ming-flash-omni-2.0 has no overall index value yet because there are too few independent benchmark results.

PopularityRank 229 of 875

42out of 100+1 vs. last month

Summary

Ming-flash-omni-2.0 combines image editing, video conversations, and controllable audio output in one model. InclusionAI’s omnimodal checkpoint accepts text, images, video, and audio, and generates text, images, and audio.

Its MoE architecture activates 6 of its 100 billion parameters. InclusionAI released the weights on February 11, 2026 under MIT.

All benchmark results

Every published result with its source, test setup and comparison models.

Show all 34 published benchmark results

Measurements without matching peer values

These measurements have no matching peer values under the same test conditions. Their original values and sources remain available here.

MMStar (Show measurement, test conditions, and source)

Source value: 74.88
Score: 74.88
Metric: MMStar
Unit: %
Category: source-specific
Direction: Higher is better
Source type: vendor-reported
Evaluator: HF model-card benchmark image linked to technical report
Status: active
Retrieved at: 2026-10-04
Methodology: Vendor-reported result. Multimodal/audio/video/tool-use/generation comparison table; no single independent evaluator cohort should be inferred.
Context: Source-specific observation; it is not a shared comparison cohort. Provider card; image footnote mixes report values with marked public-API/local-deployment rows and specialist output models.

Source HF model-card benchmark image linked to technical report

MMBench_EN_test11 (Show measurement, test conditions, and source)

Source value: 87
Score: 87
Metric: MMBench_EN_test11
Unit: %
Category: source-specific
Direction: Higher is better
Source type: vendor-reported
Evaluator: HF model-card benchmark image linked to technical report
Status: active
Retrieved at: 2026-10-04
Methodology: Vendor-reported result. Multimodal/audio/video/tool-use/generation comparison table; no single independent evaluator cohort should be inferred.
Context: Source-specific observation; it is not a shared comparison cohort. Provider card; image footnote mixes report values with marked public-API/local-deployment rows and specialist output models.

Source HF model-card benchmark image linked to technical report

HallusionBench (Show measurement, test conditions, and source)

Source value: 66.14
Score: 66.14
Metric: HallusionBench
Unit: %
Category: source-specific
Direction: Higher is better
Source type: vendor-reported
Evaluator: HF model-card benchmark image linked to technical report
Status: active
Retrieved at: 2026-10-04
Methodology: Vendor-reported result. Multimodal/audio/video/tool-use/generation comparison table; no single independent evaluator cohort should be inferred.
Context: Source-specific observation; it is not a shared comparison cohort. Provider card; image footnote mixes report values with marked public-API/local-deployment rows and specialist output models.

Source HF model-card benchmark image linked to technical report

MMVet (Show measurement, test conditions, and source)

Source value: 85.64
Score: 85.64
Metric: MMVet
Unit: %
Category: source-specific
Direction: Higher is better
Source type: vendor-reported
Evaluator: HF model-card benchmark image linked to technical report
Status: active
Retrieved at: 2026-10-04
Methodology: Vendor-reported result. Multimodal/audio/video/tool-use/generation comparison table; no single independent evaluator cohort should be inferred.
Context: Source-specific observation; it is not a shared comparison cohort. Provider card; image footnote mixes report values with marked public-API/local-deployment rows and specialist output models.

Source HF model-card benchmark image linked to technical report

MMMU (Show measurement, test conditions, and source)

Source value: 77.11
Score: 77.11
Metric: MMMU
Unit: %
Category: source-specific
Direction: Higher is better
Source type: vendor-reported
Evaluator: HF model-card benchmark image linked to technical report
Status: active
Retrieved at: 2026-10-04
Methodology: Vendor-reported result. Multimodal/audio/video/tool-use/generation comparison table; no single independent evaluator cohort should be inferred.
Context: Source-specific observation; it is not a shared comparison cohort. Provider card; image footnote mixes report values with marked public-API/local-deployment rows and specialist output models.

Source HF model-card benchmark image linked to technical report

MVbench (Show measurement, test conditions, and source)

Source value: 77.5
Score: 77.5
Metric: MVbench
Unit: %
Category: source-specific
Direction: Higher is better
Source type: vendor-reported
Evaluator: HF model-card benchmark image linked to technical report
Status: active
Retrieved at: 2026-10-04
Methodology: Vendor-reported result. Multimodal/audio/video/tool-use/generation comparison table; no single independent evaluator cohort should be inferred.
Context: Source-specific observation; it is not a shared comparison cohort. Provider card; image footnote mixes report values with marked public-API/local-deployment rows and specialist output models.

Source HF model-card benchmark image linked to technical report

MLVU (Show measurement, test conditions, and source)

Source value: 77.72
Score: 77.72
Metric: MLVU
Unit: %
Category: source-specific
Direction: Higher is better
Source type: vendor-reported
Evaluator: HF model-card benchmark image linked to technical report
Status: active
Retrieved at: 2026-10-04
Methodology: Vendor-reported result. Multimodal/audio/video/tool-use/generation comparison table; no single independent evaluator cohort should be inferred.
Context: Source-specific observation; it is not a shared comparison cohort. Provider card; image footnote mixes report values with marked public-API/local-deployment rows and specialist output models.

Source HF model-card benchmark image linked to technical report

librispeech_test-other (down) (Show measurement, test conditions, and source)

Source value: 2.2
Score: 2.2
Metric: librispeech_test-other (down)
Unit: %
Category: source-specific
Direction: Lower is better
Source type: vendor-reported
Evaluator: HF model-card benchmark image linked to technical report
Status: active
Retrieved at: 2026-10-04
Methodology: Vendor-reported result. Multimodal/audio/video/tool-use/generation comparison table; no single independent evaluator cohort should be inferred.
Context: Source-specific observation; it is not a shared comparison cohort. Provider card; image footnote mixes report values with marked public-API/local-deployment rows and specialist output models.

Source HF model-card benchmark image linked to technical report

AlpacaEval (Show measurement, test conditions, and source)

Source value: 95.6
Score: 95.6
Metric: AlpacaEval
Unit: %
Category: source-specific
Direction: Higher is better
Source type: vendor-reported
Evaluator: HF model-card benchmark image linked to technical report
Status: active
Retrieved at: 2026-10-04
Methodology: Vendor-reported result. Multimodal/audio/video/tool-use/generation comparison table; no single independent evaluator cohort should be inferred.
Context: Source-specific observation; it is not a shared comparison cohort. Provider card; image footnote mixes report values with marked public-API/local-deployment rows and specialist output models.

Source HF model-card benchmark image linked to technical report

ContextASR-Bench (down) (Show measurement, test conditions, and source)

Source value: 3.3
Score: 3.3
Metric: ContextASR-Bench (down)
Unit: %
Category: source-specific
Direction: Lower is better
Source type: vendor-reported
Evaluator: HF model-card benchmark image linked to technical report
Status: active
Retrieved at: 2026-10-04
Methodology: Vendor-reported result. Multimodal/audio/video/tool-use/generation comparison table; no single independent evaluator cohort should be inferred.
Context: Source-specific observation; it is not a shared comparison cohort. Provider card; image footnote mixes report values with marked public-API/local-deployment rows and specialist output models.

Source HF model-card benchmark image linked to technical report

WSC-Eval-TTS-Avg-WER (down) (Show measurement, test conditions, and source)

Source value: 2.71
Score: 2.71
Metric: WSC-Eval-TTS-Avg-WER (down)
Unit: %
Category: source-specific
Direction: Lower is better
Source type: vendor-reported
Evaluator: HF model-card benchmark image linked to technical report
Status: active
Retrieved at: 2026-10-04
Methodology: Vendor-reported result. Multimodal/audio/video/tool-use/generation comparison table; no single independent evaluator cohort should be inferred.
Context: Source-specific observation; it is not a shared comparison cohort. Provider card; image footnote mixes report values with marked public-API/local-deployment rows and specialist output models.

Source HF model-card benchmark image linked to technical report

WSC-Eval-TTS-Avg-ACC (Show measurement, test conditions, and source)

Source value: 83.25
Score: 83.25
Metric: WSC-Eval-TTS-Avg-ACC
Unit: %
Category: source-specific
Direction: Higher is better
Source type: vendor-reported
Evaluator: HF model-card benchmark image linked to technical report
Status: active
Retrieved at: 2026-10-04
Methodology: Vendor-reported result. Multimodal/audio/video/tool-use/generation comparison table; no single independent evaluator cohort should be inferred.
Context: Source-specific observation; it is not a shared comparison cohort. Provider card; image footnote mixes report values with marked public-API/local-deployment rows and specialist output models.

Source HF model-card benchmark image linked to technical report

Page 1 of 3

Popularity compared

How often the model is searched for, used and mentioned, compared with every model of its category.

Popularity index

The 100 most popular models of this category, 0 to 100, higher is more popular

20 of 101 models

GPT-573

DeepSeek V4.1 Flash 101768

Kimi K368

DiffusionGemma 26B A4B IT63

Claude Opus 562

Qwen 3.8 Flash62

Ming-flash-omni-2.042

DeepSeek V4.1 Flash 1017

DiffusionGemma 26B A4B IT

Claude Opus 5

Qwen 3.8 Flash

Ming-flash-omni-2.0

As of November 2026 · Source: gradually.ai popularity index

Over time

Weekly values from October 18, 2026 to November 5, 2026

Ming-flash-omni-2.0GLM-4.5VLlama 3.2 3BOrnith 1.0 9BOrnith 1.5 35B A3B

All specifications

Architecture

Parameters: 100 billion total, 6 billion active; Source: Parameters

Input and output

Modalities: Input: image, text, video, and audio. Output: image, text, and audio.; Source: Modalities

Watermarking and provenance

AI watermarking refers to invisible signals embedded in the output. Visible marks and provenance metadata are listed separately. Unknown means there is insufficient evidence for that particular field.

AI watermarking: Unknown; no source
Watermarking technique: Not known; no source

Licence and access

Release: February 11, 2026; Source: Release
License: MIT; Source: License

Frequently asked questions about Ming-flash-omni-2.0

Can I run Ming-flash-omni-2.0 locally?

Yes, the weights are open. With 100B parameters it needs a lot of memory, more a server than a laptop.

Stated by InclusionAI

Open weights: yes
License: MIT
Parameters: 100B
active per token: 6B

Sources and data date

Show all 10 sources

Benchmark sources at a glance

Which independent evaluators tested this model. This is coverage, not a quality ranking.

No active result from the reviewed external evaluators can currently be attributed to this model. Other source evidence remains available below.

Without vendor claims and historical or disputed results; those stay labelled in the single values.

Type	Evidence and data date
Research status	Research date October 4, 2026. 4 source URLs checked. This documents the inspected sources, not an exhaustive inventory of every publication.
Not published	sample-count-and-confidence-interval Original wording The paper describes the 2,200-question corpus but publishes neither a per-model denominator nor confidence intervals for these macro-averages. Corpus total is not assigned as a per-score sampleCount.
Unresolved	checkpoint-and-inference-mode Original wording The paper names Ming-Flash-Omni-2.0 and says open models follow official inference configurations, but does not identify an immutable weight revision, exact prompt/configuration, or reasoning effort.
Unresolved	per-cell sample count and confidence interval; immutable checkpoint revision; inference route Original wording Not stated in the inspected Table 2 cell context; left unknown.
Additional source	Ming-flash-omni 2.0 model card (retrieved October 3, 2026; October 4, 2026) · Ming-flash-omni 2.0 model card · Editorial description reviewed October 4, 2026
Additional source	Feng et al., TraceAV-Bench, arXiv (retrieved October 17, 2026)
Additional source	TraceAV-Bench arXiv paper v2, Table 2 and Table 7 (retrieved October 17, 2026; October 18, 2026)
Additional source	Tao et al., LVOmniBench paper (arXiv:2603.19217) (retrieved October 22, 2026)
Additional source	OpenRouter: Popularity index (derived), CC BY 4.0
Additional source	Vercel AI Gateway: Popularity index (derived), CC BY 4.0
Additional source	Ollama: cumulative downloads for the popularity index
Additional source	Replicate: cumulative runs for the popularity index
Additional source	Google Trending Now: limited feed selection for Germany and the US
Additional source	Creative Commons Attribution 4.0